Plotting ROC curve in R Programming

Last Updated : 07 Oct, 2024

In this article, we will discuss what is ROC curve, the Importance of ROC Curves in Model Evaluation, and how to plot the ROC curve in R Programming Language.

Introduction to ROC Curve in R

In binary classification problems, evaluating the performance of a model is crucial. One common method to assess a model's classification performance is through the Receiver Operating Characteristic (ROC) curve. The ROC curve in R is a graphical representation that explains the diagnostic ability of a binary classifier system as its discrimination threshold is varied. It plots the True Positive Rate (Sensitivity) against the False Positive Rate (1 - Specificity) at various threshold settings.

True Positive Rate (TPR): The proportion of actual positives correctly identified by the model (also known as sensitivity or recall).
False Positive Rate (FPR): The proportion of actual negatives incorrectly identified as positives by the model.
Area Under the Curve (AUC): The area under the ROC curve, quantifies the overall ability of the model to discriminate between positive and negative classes. A perfect model has an AUC of 1, while a random guess has an AUC of 0.5.

Importance of ROC Curves in Model Evaluation

The ROC curve in R helps in understanding how well the model performs across different thresholds. It provides a visual understanding of the trade-off between true positives and false positives. The ROC curve is particularly helpful when:

Imbalanced dataset where one class dominates over the other.
Compare the performance of multiple classification models.
Interested in how the classifier performs over a range of thresholds.

To plot an ROC curve in R, you will need some commonly used libraries, such as pROC and ROCR. Here's how to install them:

library(pROC) library(ROCR)

1: Plotting ROC Curve Using `pROC`

The pROC package makes it simple to compute and visualize ROC curves. Let's start with a basic example using a simulated dataset.

Generate or load data: Assume we have a binary outcome (0 or 1) and corresponding predicted probabilities from a logistic regression model.
Use roc() function to calculate the ROC curve.
Plot the ROC curve and compute the Area Under the Curve (AUC).

# Simulated data set.seed(123) actual <- sample(c(0, 1), 100, replace = TRUE)  # Actual binary outcomes predicted_probs <- runif(100)  # Predicted probabilities from the model  # Load pROC package library(pROC)  # Calculate ROC curve roc_curve <- roc(actual, predicted_probs)  # Plot ROC curve plot(roc_curve, col = "blue", main = "ROC Curve", print.auc = TRUE)  # Add additional information to the plot abline(a = 0, b = 1, lty = 2, col = "red")  # Diagonal line for random guess

Output:

In this example, we used roc() to calculate the ROC curve and plot() to visualize it. The diagonal red line indicates a random classifier. The ROC curve is plotted, and the AUC value is displayed on the plot.

2: Plotting ROC Curve Using `ROCR`

The ROCR package offers flexibility in terms of plotting and evaluating the ROC curve with more customizable options.

Use the prediction() function to generate the ROC curve data.
Calculate performance metrics like TPR and FPR.
Plot the ROC curve.

# Load ROCR package library(ROCR)  # Create prediction object pred <- prediction(predicted_probs, actual)  # Calculate performance (True Positive Rate and False Positive Rate) perf <- performance(pred, "tpr", "fpr")  # Plot the ROC curve plot(perf, col = "darkgreen", lwd = 2, main = "ROC Curve with ROCR")  # Add a diagonal reference line abline(a = 0, b = 1, col = "red", lty = 2)

Output:

This code plots the ROC curve using ROCR and adds a diagonal reference line representing random guessing.

High AUC value (close to 1): This indicates that the model has a high ability to distinguish between positive and negative classes.
Low AUC value (close to 0.5): This suggests the model performs no better than random guessing.
ROC curve closer to the top-left corner: A good-performing model that has high sensitivity and specificity.

Conclusion

The ROC curve in R is an essential tool for evaluating the performance of binary classification models. In R, packages like pROC and ROCR offer efficient ways to compute and visualize ROC curves. By understanding the trade-off between true positives and false positives, as well as calculating the Area Under the Curve (AUC), you can make informed decisions on the performance of your models.

Plotting ROC curve in R Programming

therebootedcoder

Improve

Article Tags :

Plotting ROC curve in R Programming

Introduction to ROC Curve in R

Importance of ROC Curves in Model Evaluation

1: Plotting ROC Curve Using pROC

2: Plotting ROC Curve Using ROCR

Conclusion

Similar Reads

1: Plotting ROC Curve Using `pROC`

2: Plotting ROC Curve Using `ROCR`